Study on Web Templates
From November 1 through November 15, we are conducting a study
regarding the feasibility of sharing document templates between
different web applications.
We need your help
For that study, we have collected a number of example programs at
the WASH/CGI example gallery
and all we ask for is ten minutes of your time to try out as many
of the example programs as possible. The more buttons you click
and the more forms you fill out in different example programs, the
better.
Each output page that you see on the screen is saved for later evaluation.
Your Reward
From time to time (roughly, on every 32nd screen you visit), a pop up
window will appear where
you can register your email address for a price draw for 1x50 EURO
and 2x25 EURO. Each email address can win at most one price. The
prices cannot be attained by legal process. The price draw closes
on November 15. The winners will be
notified by 30.11.2003 and their names will be posted on
http://proglang.informatik.uni-freiburg.de.
The winners
My apologies that the selection of the winners and their notification
has taken until 08.12.2003. Their names are
- Sreedhar Mallela
- Harald Hiss
- Sasche Ysop
Tell your friends
...the more data we collect, the better.
Background of the study
We are currently extending the Web programming language
WASH with a mechanism for caching
fragments of HTML pages. The basic ideas are described in
a paper on BigWig and
a paper on
WASH. The goal of the study is twofold.
-
First, we want to collect "real" data on the effectiveness of
template-based caching.
-
Second, we want to explore the possibility of sharing
templates among different applications. (that is, how many
templates are used over and over in many applications)
Thank you for your help
What happened ...
-
On Friday Oct 31, one hour after I sent the announcement, I
realized that only one script was properly templatified. Two
hours and more than 6000 page views later, all scripts were
replaced by templatified versions.
-
On Wednesday Nov 5, there was a report of intermittent
crashing that was not readily reproducible. The problem was
that the timeout of the locking procedure was too short. The
solution was to change the test for the timeout to also check
the modification time of the locked resource.
-
On Wednesday Nov 5, everything stopped working in the late
afternoon. The reason was that the logger had created a
directory with 31999 entries (that is 31999 page views), which
is apparently the maximum on nakalele's file system. After
changing the directory structure, all scripts were back
online at 20:30.
-
There was a bug in GuessNumberNoCheating that caused a guess
of 100 to be rejected as ``Number too small´´.
Peter Thiemann
Last modified: Mon Mar 29 11:23:29 CEST 2004